INDEX

Explanations

phrases related to doses, especially related to medication

oai_token-act-pair · gpt-3.5-turbo

references to dosage in medical or pharmaceutical contexts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

inates

-0.75

 Yard

-0.66

atters

-0.65

 Borders

-0.65

¬

-0.63

aldi

-0.63

penter

-0.62

organized

-0.62

American

-0.62

ovies

-0.62

POSITIVE LOGITS

 dose

1.15

 doses

0.91

 regimen

0.85

 dosage

0.85

 aspirin

0.82

 inhal

0.77

 orally

0.77

 nicotine

0.76

 weights

0.76

 poisoning

0.74

Activations Density 0.017%

phrases related to doses, especially related to medication

references to dosage in medical or pharmaceutical contexts

No Comments

No Known Activations